[Rocks-Discuss] rocks 6.1 kickstart compute nodes stops at language screen

Dimitris Mavroeidis dmavroeidis at csri.gr
Fri Jun 7 07:23:05 PDT 2013


Hi Marcin,

I have the same problem after upgrading the CentOS packages. My compute 
nodes cannot pxe boot.

Apache seems to be working, since `rocks list host profile compute-0-0` 
does not give me an error.

Though, when running "wget --no-check-certificate 
https://localhost/install/sbin/public/kickstart.cgi", I get an empty file.

I have tried several suggestions from previous posts, but nothing seems 
to work.

If you finally come up with a solution, please share it with the list.

Best regards,
Dimitris



On 31/05/2013 10:56 μμ, Marcin Sliwowski wrote:
> Luca,
>
> I have the following rolls installed
>
> ganglia:      6.1     x86_64 yes
> web-server:   6.1     x86_64 yes
> perl:         6.1     x86_64 yes
> kernel:       6.1     x86_64 yes
> python:       6.1     x86_64 yes
> area51:       6.1     x86_64 yes
> updates:      6.4     x86_64 yes
> base:         6.1     x86_64 yes
> java:         6.1     x86_64 yes
> os:           6.4     x86_64 yes
> service-pack: 6.1     x86_64 yes
>
> I think you are right about apache not being able to connect to the 
> DB, as apache a `rocks list host profile compute-0-0` dumps that exact 
> error.
>
> Now to try and figure out what exactly is breaking it.
>
> Thanks
>
> Marcin Sliwowski | SysAdmin at RENCI | (919) 445-0479
>
> On 05/31/2013 03:43 PM, Luca Clementi wrote:
>> On Fri, May 31, 2013 at 10:05 AM, Marcin Sliwowski <marcin at renci.org> 
>> wrote:
>>> I am pretty confidant that httpd is still functional and doing its 
>>> job. It
>>> does log both the attempts made and the errors involved in pulling 
>>> down the
>>> kickstart file.
>>>
>>> I thought that maybe someone would recognize the error that is being 
>>> logged
>>> in /var/log/httpd/ssl_error_log:
>>>
>>> Wed May 29 14:32:02 2013] [error] [client 127.0.0.1] Traceback (most 
>>> recent
>>> call last):
>>> [Wed May 29 14:32:02 2013] [error] [client 127.0.0.1]   File
>>> "/opt/rocks/bin/rocks", line 301, in <module>
>>> [Wed May 29 14:32:02 2013] [error] [client 127.0.0.1]   File
>>> "/opt/rocks/bin/rocks", line 301, in <module>
>>> [Wed May 29 14:32:02 2013] [error] [client 127.0.0.1]
>>> command.runWrapper(name, args[i:])
>>> [Wed May 29 14:32:02 2013] [error] [client 127.0.0.1]   File
>>> "/opt/rocks/lib/python2.6/site-packages/rocks/commands/__init__.py", 
>>> line
>>> 2194, in runWrapper
>>> [Wed May 29 14:32:02 2013] [error] [client 127.0.0.1] 
>>> self.run(self._params,
>>> self._args)
>>> [Wed May 29 14:32:02 2013] [error] [client 127.0.0.1]   File
>>> "/opt/rocks/lib/python2.6/site-packages/rocks/commands/list/host/attr/__init__.py", 
>>>
>>> line 129, in run
>>> [Wed May 29 14:32:02 2013] [error] [client 127.0.0.1]     for host in
>>> self.getHostnames(args):
>>> [Wed May 29 14:32:02 2013] [error] [client 127.0.0.1]   File
>>> "/opt/rocks/lib/python2.6/site-packages/rocks/commands/__init__.py", 
>>> line
>>> 773, in getHostnames
>>> [Wed May 29 14:32:02 2013] [error] [client 127.0.0.1] min,max =
>>> self.db.fetchone()
>>> [Wed May 29 14:32:02 2013] [error] [client 127.0.0.1] TypeError: 
>>> 'NoneType'
>>> object is not iterable
>>>
>> I would say that apache can not connect anymore to the rocks mysql DB.
>> It is difficult to say why though...
>>
>> Which roll do you have installed? (rocks list roll)
>>
>> You could try to enable the apache user, so you can "su - apache", and
>> then try to run some "rocks list" command.
>>
>> rocks list host profile compute-0-0-0
>>
>> I suppose that some update to mysql/mysql-python/mysql-libs changed
>> something which broke this.
>>
>> Luca
>

-------------- next part --------------
A non-text attachment was scrubbed...
Name: dmavroeidis.vcf
Type: text/x-vcard
Size: 385 bytes
Desc: not available
Url : https://lists.sdsc.edu/pipermail/npaci-rocks-discussion/attachments/20130607/3ab890b0/dmavroeidis.vcf 


More information about the npaci-rocks-discussion mailing list